VZFLO AI

Use Datasets for Workspace Tables

Datasets are editable workspace tables for structured records such as vendor lists, quote queues, payroll totals, inventory notes, or customer handoff data. Workspace agents can read datasets, and when allowed, write dataset records as part of their automation work.

Step by step

1

Open Datasets

Sign in to VZFLO and open Datasets from the app navigation. The Datasets page lists every table in the current workspace.

VZFLO / Datasets

Datasets

Structured tables that people and workspace agents can use.

New dataset
Name Description Rows
Vendor Directory Approved payees and payment links 12

Open Datasets to view or create workspace tables.

Use a dataset when the information belongs to the workspace and should be reused by people or workspace agents later.

2

Create a dataset

Choose New dataset, give it a short name, and add an optional description that explains what kind of records belong there.

VZFLO / New Dataset
Create dataset

Name the table for the job it performs in the workspace.

Good names are operational: Vendor Directory, Quote Requests, Payroll Review, Inventory Watchlist, or Escalation Rules.

3

Edit the grid

Rename columns in the header row, add rows or columns as needed, and type values into cells. Changes autosave as you work.

VZFLO / Vendor Directory
Saved Add column Add row

The grid autosaves cell edits as you work.

Use normal columns for business data like Payee, URL, Status, Owner, Amount, or Notes.

4

Understand how records are stored

Each dataset has column records, row records, and cell records. A row becomes useful data when cells in that row contain values.

Dataset storage model
RowsDurable row IDs and display order.
ColumnsNames, order, and data type.
CellsOne value for one row and one column.
row_id + column_id -> value
54b2...8f91 + Payee -> Mesa Supplies
54b2...8f91 + Status -> Approved

Datasets store structured rows, columns, and cell values instead of one large note.

Empty cells are not meaningful records. If a workspace agent needs to find something, give it clear column names and consistent values.

5

Allow workspace agents to read datasets

When creating or editing a workspace agent, turn on Datasets under Read access. This lets the agent inspect dataset names, columns, rows, and values.

VZFLO / Read access

Check Datasets when an agent needs table lookup ability.

Read access is enough for lookup tasks like checking approved vendors, reading escalation rules, or summarizing open rows.

6

Allow workspace agents to update datasets

If the agent should add or change records, also turn on Dataset records under Write access. Save the agent after changing permissions.

VZFLO / Write access

Enable Dataset records only for agents that should modify table data.

Only enable write access for agents with narrow instructions, such as “add one row for each qualified quote request” or “update the Status column only.”

Launch checklist

  • The dataset has a clear name and description.
  • Column names describe the data the team or agent should use.
  • Important rows have values in normal business columns.
  • Workspace agents that need lookup ability have Datasets read access.
  • Workspace agents that should modify records have Dataset records write access.
  • Agent instructions say which dataset, columns, and update rules to use.

Troubleshooting

A workspace agent says it cannot see dataset data.

Open the agent settings and enable Datasets under Read access, then rerun the agent.

A workspace agent can read the dataset but cannot update rows.

Enable Dataset records under Write access and make sure the agent instructions explain exactly what it may change.

The dataset is hard for agents to use.

Rename vague columns, avoid mixing multiple record types in one table, and keep values consistent across rows.

Datasets · 10 min

Connect an HVAC Service Customer Dataset to a Voice Agent

Create an HVAC service customer list, attach it to a voice agent, and let the assistant look up or create tentative service dates.